terraform-provider-kubernetes: v2.0.1: Resources cannot be created. Does kubectl refference to kube config properly?

Terraform version: v0.14.4
Kubernetes provider version: v2.0.1
Helm provider version: v1.3.2

Steps to Reproduce

I use a GitLab pipeline to deploy helm charts on my Kubernetes cluster by using the helm terraform provider.

provider "helm" {
  kubernetes {
    config_path = "~/.kube/config"
  }
}

Since version v2.0.1 of the Kubernetes provider the Helm provider is not able to access to the kube config file properly. The error message looks like:

module.helm.helm_release.nginx-ingress-internal: Creating...
Error: configmaps is forbidden: User "system:serviceaccount:gitlab-prod:default" cannot create resource "configmaps" in API group "" in the namespace "nginx-ingress"
Error: namespaces is forbidden: User "system:serviceaccount:gitlab-prod:default" cannot create resource "namespaces" in API group "" at the cluster scope
Error: namespaces is forbidden: User "system:serviceaccount:gitlab-prod:default" cannot create resource "namespaces" in API group "" at the cluster scope
Error: namespaces is forbidden: User "system:serviceaccount:gitlab-prod:default" cannot create resource "namespaces" in API group "" at the cluster scope
Error: namespaces is forbidden: User "system:serviceaccount:gitlab-prod:default" cannot create resource "namespaces" in API group "" at the cluster scope
Error: namespaces is forbidden: User "system:serviceaccount:gitlab-prod:default" cannot create resource "namespaces" in API group "" at the cluster scope

The reason why I use Helm provider v1.3.2 is described in this bug report: https://github.com/hashicorp/terraform-provider-helm/issues/662

Temporary solution

Revert back to version v1.13.3

Community Note

Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
If you are interested in working on this issue or have submitted a pull request, please leave a comment

About this issue

Original URL
State: closed
Created 3 years ago
Reactions: 3
Comments: 15 (7 by maintainers)

Most upvoted comments

@jrhouston you said you didn’t change “anything about the way the kubeconfig gets loaded”. But the changelog says something different:

2.0.0 (January 21, 2021)

BREAKING CHANGES:

Remove default of ~/.kube/config for config_path (#1052)

Honestly I don’t understand that. ~/.kube/config is the standard! So why removing a standard that everyone is actually using?

tantweiler on Jan 25, 2021

@tantweiler In your example I see your provider kubernetes block is empty, but your provider helm block has a config_path set. You need to set it in both provider blocks as both providers need to know the path to the kubeconfig. Did you try that?

jrhouston on Jan 24, 2021